Introduction
Figure 2: Schematic illustration of the circle estimation process.
Figure 1: Schematic overview of the proposed approach
This work proposes a novel tree detection methodology, named DTCD (Distance Transform Circle Detection), based on a fast circle detection method via Distance Transform and Akaike Information Criterion (AIC) optimization [1]. More specifically, a visible-band vegetation index (RGBVI) is calculated to enhance canopy regions, followed by morphological filtering to delineate individual tree crowns. The Euclidean Distance Transform is then applied, and the local maxima of the smoothed distance map are extracted as candidate tree locations. The final detections are iteratively refined using the AIC to optimize the number of trees with respect to canopy coverage efficiency. Additionally, this work introduces a modified tree detection algorithm using point clouds, improving the detection accuracy.
Download AgiosNikolaos-3 dataset
You can download the AgiosNikolaos-3 dataset proposed in [1] from
Google Drive: https://drive.google.com/file/d/1SNz9oOGagRoEC3Nmze4tz1SJ51WnqhKH/view?usp=sharing
Kaggle: https://www.kaggle.com/datasets/cpanag/agiosnikolaos-3
See the corresponding readme.txt files for more details.
Experiments - Downloads (available after acceptance of our Remote Sensing paper)
You can download the matlab code of the DTCD method proposed in [1]
Experimental results of DTCD on Acacia-6 dataset
Experimental results of DTCD-PC on AgiosNikolaos-3 dataset
See the corresponding readme.txt files for more details.
Introduction
This work [7] proposes an automated, unsupervised framework for 3D Spatial Tree Risk Assessment in complex urban environments that strategically integrates established point cloud analysis techniques. Building on our previously published DTCD-PC methodology, originally developed for automated tree detection, this study expands its algorithmic core to achieve a robust, unsupervised segmentation of ground, vegetation, and urban infrastructure. The framework then employs 3D voxelization, followed by K-Dimensional Tree (KD-Tree) spatial queries to quantitatively evaluate structural proximity and road clearance violations.
Download AgiosNikolaos-3D dataset
You can download the AgiosNikolaos-3D dataset proposed in [7] from
Google Drive: https://drive.google.com/drive/folders/1dJ0hSjDQh6t9ypF1ZwF7xmbbw6W069YY?usp=sharing
Kaggle: https://www.kaggle.com/datasets/cpanag/agiosnikolaos-3d
See the corresponding readme.txt files for more details.
Experiments - Downloads (available after acceptance of our Remote Sensing paper)
You can download the matlab code of the method proposed in [7]
Experimental results on AgiosNikolaos-3 dataset
See the corresponding readme.txt files for more details.
Related Publications (Tree Detection and Urban Tree Risk Assessment)
[1] S. Markaki and C. Panagiotakis, Unsupervised Tree Detection from UAV Imagery and 3D Point Clouds via Distance Transform-Based Circle Estimation and AIC Optimization, Remote Sensing. 18(3):505, 2026.
[2] S. Markaki, C. Panagiotakis, Unsupervised Tree Detection and Counting via Region-based Circle Fitting, International Conference on Pattern Recognition Applications and Methods (ICPRAM), 2023.
[3] C. Panagiotakis and A. Argyros, Parameter-free Modelling of 2D Shapes with Ellipses, Pattern Recognition, vol. 53, pp. 259-275, 2016.
[4] Tong, P., Han, P., Li, S., Li, N., Bu, S., Li, Q., and Li, K. (2021). Counting trees with point-wise supervised segmentation network. Engineering Applications of
Artificial Intelligence, 100:104172
[5] C. Panagiotakis and A. Argyros, Cell Segmentation via Region-based Ellipse Fitting, IEEE International Conference on Image Processing, 2018.
[6] C. Panagiotakis and A.A. Argyros, Region-based Fitting of Overlapping Ellipses and its Application to Cells Segmentation, Image and Vision Computing, Elsevier, vol. 93, pp. 103810, 2020.
[7] S. Markaki and C. Panagiotakis, Urban Tree Risk Assessment: A Voxel-based Spatial Analysis Framework using Photogrammetric Point Clouds, Remote Sensing (under review). 2026. Preprints: